The dynamic response of splash erosion to aggregate mechanical breakdown through rainfall simulation events in Ultisols (subtropical China)

Rainfall splash erosion is a temporal dynamic process that is determined by the interaction of soil surface and rainfall conditions. The properties of aggregate, such as aggregate stability and size distribution, are recognized as factors that influence the variation in soil surface structure and sediment selectivity during splash erosion event. This study based on aggregate stability to investigate the effect of aggregates breakdown dynamics on splash erosion and sediment size selectivity processes of Ultisols in subtropical China during rainfall events. Laboratory experiments of rainfall simulation involving rainstorm simulations at 58.1 +/- 2.8 mm h(-1) were conducted on four soils derived from Quarternary red clay and Shale. We compared the size distributions of aggregates produced by raindrop impact with the sediment size distributions resulted from transportation. Fractal dimension (D) and enrichment ratio (ER) were used to characterize aggregates size distribution and sediment size selectivity. Aggregates were prone to breakdown at the beginning of rainfall and a multiple regression equation (D = 1.02RMI + 1.77 T-0.06 (R-2 = 0.95, P < 0.01, n = 60)) was obtained to describe aggregate breakdown dynamics considering a relative aggregate stability index (RMI) and rainfall time (T). Total masses of erosion sediment decreased as a power function with an increasing size fraction. The behaviors of sediment size fractions (<0.25 mm, 0.5-0.25 mm, 1-0.5 mm, 2-1 mm, 5-2 mm) showed that the transport selectivity was decided by the interaction of aggregates size distributions and the characteristics of soil surface structure. Sediment size selectivity led to the enrichment of 1-0.25 mm fraction compared with the breakdown products. This research contributes to knowledge about the dynamics of the splash mechanism, and the results could be valuable for the development of rainfall erosion models. (C) 2014 Elsevier B.V. All rights reserved.